JULY 2006 VOLUME 19, NUMBER 7 © Copyright 2006 by the American Chemical Society
PerspectiVe Future of ToxicologysMetabolic Activation and Drug Design: Challenges and Opportunities in Chemical Toxicology Thomas A. Baillie* Department of Drug Metabolism, Merck Research Laboratories, WP75B-330, 770 Sumneytown Pike, P.O. Box 4, West Point, PennsylVania 19486-0004 ReceiVed March 20, 2006
The issue of chemically reactive drug metabolites is one of growing concern in the pharmaceutical industry inasmuch as some, but not all, reactive intermediates are believed to play a role as mediators of drug-induced toxicities. While it is now relatively straightforward to identify these short-lived electrophilic species through appropriate in vitro “trapping” experiments, our current understanding of mechanistic aspects of xenobiotic-induced toxicities is such that we cannot predict which reactive intermediates are likely to cause a toxic insult and which will be benign. Little is known about the identities of the macromolecular targets (primarily proteins) of these electrophiles or the functional consequences of their covalent modification by reactive drug metabolites. As a result, several companies have adopted approaches to minimize the potential for metabolic activation of drug candidates at the discovery/lead optimization phase as a default strategy. However, research leading to a deeper insight into mechanistic aspects of toxicities caused by reactive drug metabolites will aid greatly in the rational design of drug candidates with superior safety profiles and represents a challenging and exciting opportunity for chemical toxicology. Contents Introduction Chemically Reactive Drug Metabolites Covalent Binding to Proteins Conclusions
889 890 891 892
Introduction Today’s pharmaceutical industry is in the midst of considerable change, as companies struggle with the challenges associated with heightened global competition, patent expirations, * To whom correspondence should be addressed. Tel: 215-652-5326. Fax: 215-652-9427. E-mail:
[email protected].
increasingly complex regulatory hurdles, and societal pressures for new and effective medicines that bring true value to both patient and payer. Moreover, the general public has little appreciation of risk/benefit considerations, and the need to ensure new product safety has never been greater. The number of new chemical entities gaining marketing approval has fallen in recent years, although it has been argued that this trend is not necessarily indicative of a decrease in innovative practices within the industry (1). As a result of these pressures, the decision-making process in drug discovery and lead optimization at many companies has changed fundamentally over recent years (2). Much attention has been focused on the reasons for the continued high attrition rates in drug development, and the latest published data on this topic suggest that the primary causes are lack of efficacy in humans and preclinical toxicity (3). In the
10.1021/tx060062o CCC: $33.50 © 2006 American Chemical Society Published on Web 05/28/2006
890 Chem. Res. Toxicol., Vol. 19, No. 7, 2006
former case, the lack of appropriate animal models for novel human disease targets undoubtedly plays a role, underscoring the importance of rapid progression to clinical trials for proofof-concept in man. In the area of preclinical toxicology, the relatively high incidence of drug candidates that exhibit unacceptable toxicity profiles in animals, and therefore never advance to human trials, highlights the critical importance of safety assessment in pharmaceutical research and development and the need for enhanced predictive capabilities in support of drug candidate selection. It is recognized that established approaches to preclinical safety evaluation are imperfect and, in some instances, have a relatively poor track record of forecasting human adverse reactions (4). Unfortunately, advances in the field of comparative toxicology have been hampered, in part, not only by the reluctance of regulatory agencies to consider novel approaches that lack extensive validation but also by the current lack of emphasis on toxicology in general as a scientific discipline. Recent developments, such as publication of the Critical Path Initiative by the U.S. Food and Drug Administration (5), should help to address the former concern, especially in the areas of biomarkers of toxicity and toxicogenomics. Nevertheless, the proprietary nature of industrial toxicology data remains a barrier to progress in defining molecular structuretoxicity relationships that could be of benefit to the field as a whole.
Chemically Reactive Drug Metabolites One of the areas that offers promise for decreasing the high rate of attrition during early drug development is that of chemically reactive drug metabolites. Because it is now widely appreciated that reactive drug metabolites, as opposed to the parent molecules from which they are derived, may serve as mediators of drug-induced toxicities, several pharmaceutical companies have implemented procedures to identify structural moieties that are subject to metabolic activation (preferably prior to entry into development) and, through appropriate structural modification, to minimize the potential for bioactivation to reactive electrophiles (6, 7). Typically, these efforts involve a combination of in vitro and in vivo studies in which reactive metabolites are identified indirectly through structural characterization of the stable end products that are formed upon capture by nucleophilic scavengers. S-Linked conjugates with the tripeptide glutathione (GSH), excreted in the bile of animals dosed with the drug candidate or formed in vitro by “trapping” techniques (normally conducted with liver preparations), have been particularly informative in this regard, although it should be recognized that certain classes of electrophilic metabolites do not form stable GSH adducts and thus may escape detection by this approach. Reactive iminium species, for example, which may be generated through metabolic oxidation of cyclic and acyclic tertiary amines, are best trapped in vitro by the addition of cyanide ions to the incubation medium, resulting in the formation of stable R-cyanoamines. Advances in liquid chromatography-tandem mass spectrometry (LC-MS/MS) techniques, notably those that afford accurate mass capabilities for assignment of the elemental compositions of parent and fragment ions, together with high-field nuclear magnetic resonance spectroscopy analysis of purified adducts, have greatly facilitated structural investigations of this type and now are employed widely by industrial Drug Metabolism and Safety Assessment groups. It should be noted that while much is already known about pathways of metabolic activation and the functional groups that are involved, much remains to be learned in this field. The never ending efforts of medicinal chemists to introduce structural novelty into drug candidates for reasons related, in part, to patent
Baillie
protection continue to present fresh opportunities for unexpected bioactivation reactions. Recent examples from our laboratories at Merck include metabolic activation of substituted piperidine (8), piperazine (9), pyrazinone (10), thiazolidinedione (11), diaminopyridine (12), and dihydrobenzoxathiin ring systems (13), the mechanisms of which were elucidated by means of the above approaches. While we have become relatively skilled in recent years in detecting and identifying reactive metabolites of potential drug candidates, advances in the field of biochemical toxicology have not kept pace with those in drug metabolism and bioanalytical chemistry, such that we are unable to predict, a priori, which reactive intermediates may be toxic to the host cell vs those that will be relatively benign. This represents a serious gap in our knowledge base, which effectively precludes meaningful predictions of the toxicological consequences of metabolic activation of a given drug candidate, since it is well-established that while some reactive intermediates appear to be mediators of end-organ toxicity, others (even if structurally very similar) are not. What, then, are the features that distinguish the two groups? There is no doubt that intrinsic chemical reactivity (and hence effective biological lifetime) plays a role, but there are likely to be many other factors that come into play in modulating the balance between cellular damage and “detoxification” following exposure to reactive electrophiles. There has been much debate, for example, about possible differences in the biological targets (notably proteins) to which reactive intermediates become covalently bound, although in most cases their identities remain unknown (see below). Targeting of different electrophiles to different intracellular organelles may occur, as appears to be the case with the reactive metabolites of acetaminophen (4′-hydroxyacetanilide, APAP) and its positional isomer, 3′-hydroxyacetanilide (AMAP; Figure 1). Thus, evidence has been obtained that the hepatotoxic metabolite of APAP, N-acetyl-para-benzoquinone imine (NAPQI), preferentially arylates proteins in liver mitochondria, whereas the reactive metabolite of AMAP (2-acetamido-para-benzoquinone) binds primarily to targets in the cytosolic fraction of hepatocytes (14). The functional significance of these findings is unclear, but the differences in subcellular disposition and toxicity profiles of APAP and AMAP are striking. Another feature of reactive metabolite chemistry that merits attention deals with the biological “transport” of these shortlived species. Studies in mice dosed with radiolabeled APAP revealed that, in addition to liver tissue, erythrocytes contained irreversibly bound radioactivity that resulted from arylation of a specific cysteinyl thiol on hemoglobin by the reactive metabolite, NAPQI (15). Interestingly, induction experiments demonstrated that the origin of the NAPQI responsible for reaction with hemoglobin was cytochrome P450-dependent bioactivation in the liver. This finding, in turn, suggested that NAPQI, a highly reactive electrophile (16), must undergo transport from its primary site of formation (the endoplasmic reticulum of liver cells) to distant locations and raised the possibility of an endogenous “carrier” molecule. Some years later, an ipso adduct of NAPQI (Figure 1), which can be formed reversibly from the quinine imine, was identified as a candidate “delivery” form of the reactive metabolite (17), thereby demonstrating that conjugation with GSH does not necessarily result only in the detoxification of reactive electrophiles (18). The issues raised above pertaining to the chemistry and biological fate of reactive drug metabolites as determinants of toxicity remain poorly defined; yet, these are factors that are critical to our understanding of the role of metabolic activation
PerspectiVe
Chem. Res. Toxicol., Vol. 19, No. 7, 2006 891
Figure 1. Proposed CYP-mediated metabolic activation pathways for (A) APAP and (B) AMAP. GSH denotes glutathione, and RSH denotes protein or nonprotein thiol. See the text for details.
in drug-induced toxicities. This area represents an important opportunity for toxicological research with profound implications for drug design and, ultimately, patient safety.
Covalent Binding to Proteins As discussed above, many biological reactive intermediates bind covalently to proteins, both close to the locus of their formation and at distant sites. In early studies of reactive metabolites, it was assumed by many investigators that such binding would be nonselective in nature and that the toxicity associated with drugs that underwent metabolic activation was a direct consequence of modifications to key structural or functional elements of exposed cells. However, while it was apparent that the extent and anatomical location of irreversibly bound drug-protein residues correlated well with the target organ and severity of the ensuing toxicity, this relationship did not hold true in all cases. Specifically, compounds such as APAP and AMAP, when dosed to animals in such a way as to afford comparable levels of binding to liver proteins, exhibited markedly different toxicological responses: APAP was hepatotoxic, and AMAP was nonhepatotoxic (14). Furthermore, the binding of these two agents to proteins was far from nonselective, in that the reactive metabolites of both compounds exhibited a high preference for arylation of cysteinyl thiols (15, 19, 20). As noted above, selectivity of binding also was noted at the subcellular organelle level, with APAP binding preferentially to liver mitochondrial proteins and AMAP binding to cytosolic targets. Clearly, the relationships between metabolic activation, covalent binding to protein, and toxicity were considerably more complex than had first been suspected.
Little progress was made on the identities of protein targets of reactive electrophiles until relatively recently when technical advances in the field of proteomics (driven largely by developments in electrospray ionization and associated LC-MS/MS methods) provided the analytical tools necessary to examine drug-protein adducts. Indeed, progress in this area over the past 2-3 years has been remarkable, as exemplified in recent work by Dennehy et al. (21) in which exposure of nuclear and cytosolic proteosomes from HEK293 cells to two model electrophiles in vitro resulted in the identification of almost 900 adducts mapped to different cysteine residues in 539 proteins. Thus, while the identification of protein targets of reactive electrophiles remains analytically challenging, primarily due to issues of dynamic range (different proteins often varying in relative abundance by a factor of 106 or more) and less-thanstoichiometric modification, the technology now exists to address this important question. It may be anticipated, therefore, that compendia of protein targets for reactive drug metabolites, such as that established at the University of Kansas (http:// tpdb.medchem.ku.edu:8080/protein_database/search.jsp), will provide insight into protein networks whose biological functions may be modulated through covalent interactions with electrophilic xenobiotics. Of particular interest at this time are sensory networks that are known to respond to electrophiles, resulting in signals that trigger cell death, stress responses, or adaptation to stress. These include the Keap-1/Nrf-2/ARE complex, GSHS-transferases, thioredoxin, and the nuclear factor κB transcriptional regulatory system (22). In each case, alkylation or oxidation of critical thiol residues may trigger characteristic biological responses, and it is tempting to speculate that while certain electrophiles modify specific thiols on these sensor proteins that signal a protective response, others may modify different sites that disrupt cellular defense mechanisms and thereby promote toxicity. From the point of view of the pharmaceutical industry, drugprotein adducts are of considerable interest for a number of reasons. First, covalent binding of radioactivity from an appropriate 14C- or tritium-labeled analogue of the drug candidate to proteins may be taken as a quantitative measure of exposure to a reactive metabolite(s) of that agent. At Merck Research Laboratories, promising drug candidates that emerge from lead optimization efforts typically are radiolabeled and assessed for their propensity to alkylate liver microsomal proteins in vitro (from several species, including human), as well as liver and plasma proteins from rats dosed under standard conditions in vivo (6). The resulting data, coupled with an understanding of bioactivation and clearance mechanisms, then are taken into consideration as one means of discriminating between candidates that are selected for further development. It should be emphasized, however, that this relatively crude approach provides no information on protein targets for binding and does not claim to serve as a predictor of toxicity. The goal of these covalent binding studies is merely to aid in the selection of development candidates with a low propensity for bioactivation in animals and humans. Second, it is widely accepted that the covalent modification of proteins by reactive xenobiotics may, under certain conditions, form the basis of a downstream immunological response mediated by antibodies that recognize the drug moiety (23). It is also recognized that such immune-mediated toxicities can be rare, but serious, events that are not always manifest in early phase clinical trials but that become apparent only in largescale phase III trials or, even worse, postmarketing. Because there are no generally applicable animal models for the human
892 Chem. Res. Toxicol., Vol. 19, No. 7, 2006
immune system (24), toxicities of this kind are the most difficult of all to predict, and until validated approaches to the assessment of human immunogenic potential become available, it seems prudent to attempt to minimize the generation of haptenized proteins that result from the metabolic activation of drug candidates. Related to this topic is the issue of the biological fate of covalently modified proteins, about which virtually nothing is known. One would anticipate that electrophiles with a high degree of chemical reactivity would bind preferentially to intracellular proteins and, following presentation by the MHC-I complex, may cause a cell-mediated immune response. Less reactive metabolites, on the other hand, would be expected to also modify extracellular proteins, undergo presentation by the MHC-II system, and potentially trigger antibody-mediated reactions. Although extremely challenging from an analytical standpoint, sensitive LC-MS/MS technology has been developed and applied to the characterization of endogenous MHC peptides and, in principle, could be extended to address xenobiotic-modified peptides. If successful, this approach could form the basis of assessing the likelihood, using appropriate human vectors, that certain types of covalent adducts may lead to immunogenic responses in man. The implications of this line of toxicological research to the pharmaceutical industry are self-evident. Third, as pointed out above, not all reactive intermediates are captured effectively by small-molecule nucleophiles, such as GSH, and it is relatively common to observe covalent binding of drug candidates to proteins in vitro that is not attenuated significantly by the addition of trapping agents to incubation media. The question then arises as to the identities of the reactive species that are binding to protein in these cases. Do the protein adducts reflect the generation of other, more highly reactive electrophiles that do not have the opportunity to diffuse far from their site of formation but instead are captured by nucleophilic centers on nearby proteins? Might knowledge of the structures of the resulting protein adducts provide new insights into unanticipated routes of metabolic activation? As noted above, mass spectrometric approaches to this problem now are available, and it is hoped that they will be applied to this intriguing area of reactive metabolite chemistry.
Conclusions From the foregoing discussion, it is apparent that the broad topic of chemically reactive drug metabolites is one replete with opportunities for toxicological research and, moreover, that the outcome of such research is likely to have a significant impact on the development of new therapeutic agents with an enhanced safety profile. Indeed, several excellent reviews on the subject of metabolic activation and its role in drug-induced toxicities have appeared recently (25-28). Why, then, has so little progress been made on this subject since the pioneering work of Brodie, Gillette, Mitchell, and their colleagues (29) at the National Institutes of Health (NIH) more than three decades ago? One reason certainly has been the lack of appropriate analytical methodology, which only recently has been rectified by technical developments in LC-MS/MS methodology for proteomic applications. A remaining problem, however, is the lack of research support for this general field of investigation. Federal granting agencies have tended to view the issue of reactive drug metabolites as a problem for industry, since it is the pharmaceutical industry that develops drugs. However, the fast-paced, results-driven nature of industrial research in many ways is incompatible with long-term, basic research programs of the type that will be needed to make headway. This “catch 22” situation has now been exacerbated by the recent reorganization of the NIH peer review system, which, as pointed out
Baillie
by Professor Liebler in the first Future of Toxicology perspective in this series (30), has essentially eliminated toxicology-related study sections, a development with serious implications for toxicological research in the United States. Whether a solution to this crisis is the formation of academia/industry/government consortia, a mechanism for enhanced sharing of toxicology data in the public domain, or other approaches remains unclear, but the opportunities for toxicology as an integral component of contemporary pharmaceutical research have never been brighter. To comment on this and other Future of Toxicology perspectives, please visit our Perspectives Open Forum at http:// pubs.acs.org/journals/crtoec/openforum.
References (1) Schmid, E. F., and Smith, D. A. (2005) Is declining innovation in the pharmaceutical industry a myth? Drug DiscoVery Today 10, 1031-1039. (2) MacCoss, M., and Baillie, T. A. (2004) Organic chemistry in drug discovery. Science 303, 1810-1813. (3) Kola, I., and Landis, J. (2004) Can the pharmaceutical industry reduce attrition rates? Nat. ReV. Drug DiscoVery 3, 711-716. (4) Olsen, H., Betton, G., Robinson, D., Thomas, K., Munro, A., Kolaja, G., Lilly, P., Sanders, J., Sipes, G., Bracken, W., et al. (2000) Concordance of the toxicity of pharmaceuticals in humans and animals. Regul. Toxicol. Pharmacol. 32, 56-67. (5) U.S. Department of Health and Human Services, Food and Drug Administration (2004) Challenges and opportunity on the critical path to new medical products. http://www.fda.gov/oc/initiatives/criticalpath/ whitepaper.html. (6) Evans, D. C., Watt, A. P., Nicoll-Griffith, D. A., and Baillie, T. A. (2004) Drug-protein covalent adducts: an industry perspective on minimizing the potential for drug bioactivation in drug discovery and development. Chem. Res. Toxicol. 17, 3-16. (7) Kalgutkar, A. S., and Soglia, J. R. (2005) Minimising the potential for metabolic activation in drug discovery. Exp. Opin. Drug Metab. Toxicol. 1, 91-142. (8) Yin, W., Mitra, K., Stearns, R. A., Baillie, T. A., and Kumar, S. (2004) Conversion of the 2,2,6,6-tetramethylpiperidine moiety to a 2,2dimethylpyrrolidine by cytochrome P450: Evidence for a mechanism involving nitroxide radicals and heme iron. Biochemistry 43, 5455-5466. (9) Doss, G. A., Miller, R. R., Zhang, Z., Teferra, Y., Nargund, R. P., Palucki, B., Park, M. K., Tang, Y. S., Evans, D. C., Baillie, T. A., and Stearns, R. A. (2005) Metabolic activation of a 1,3-disubstituted piperazine derivative: Evidence for a novel ring contraction to an imidazoline. Chem. Res. Toxicol. 18, 271-276. (10) Singh, R., Silva-Elipe, M. V., Pearson, P. G., Arison, B. H., Wong, B. K., White, R., Yu, X., Burgey, C. S., Lin, J. H., and Baillie, T. A. (2003) Metabolic activation of a pyrazinone-containing thrombin inhibitor. Evidence for novel biotransformation involving pyrazinone ring oxidation, rearrangement and covalent binding to proteins. Chem. Res. Toxicol. 16, 198-207. (11) Reddy, V. B. G., Karanam, B. V., Gruber, W. L., Wallace, M. A., Vincent, S. H., Franklin, R. B., and Baillie, T. A. (2005) Mechanistic studies on the metabolic scission of thiazolidinedione derivatives to acyclic thiols. Chem. Res. Toxicol. 18, 880-888. (12) Tang, C., Subramanian, R., Kuo, Y., Krymgold, S., Lu, P., Kuduk, S. G., Ng, C., Feng, D.-M., Elmore, C., Soli, E., Ho, J., Bock, M. G., Baillie, T. A., and Prueksaritanont, T. (2005) Bioactivation of 2,3diaminopyridine-containing bradykinin B1 receptor antagonists: Irreversible binding to liver microsomal proteins and formation of glutathione conjugates. Chem. Res. Toxicol. 18, 934-945. (13) Zhang, Z., Chen, Q., Li, Y., Doss, G. A., Dean, B. J., Ngui, J. S., Silva-Elipe, M., Kim, S., Wu, J. Y., DiNinno, F., Hammond, M. L., Stearns, R. A., Evans, D. C., Baillie, T. A., and Tang, W. (2005) In vitro bioactivation of dihydrobenzoxathiin selective estrogen receptor modulators by cytochrome P450 3A4 in human liver microsomes: Formation of reactive iminium and quinine type metabolites. Chem. Res. Toxicol. 18, 675-685. (14) Tirmenstein, M. A., and Nelson, S. D. (1989) Subcellular binding and effects on calcium homeostasis produced by acetaminophen and a nonhepatotoxic regioisomer, 3′-hydroxy-acetanilide, in mouse liver. J. Biol. Chem. 264, 9814-9819. (15) Axworthy, D. B., Hoffmann, K.-J., Streeter, A. J., Calleman, C. J., Pascoe, G. A., and Baillie, T. A. (1988) Covalent binding of acetaminophen to mouse hemoglobin. Identification of major and minor adducts formed in vivo and implications for the nature of the arylating metabolites. Chem.-Biol. Interact. 68, 99-116.
PerspectiVe (16) Dahlin, D. C., Miwa, G. T., Lu, A. Y. H., and Nelson, S. D. (1984) N-Acetyl-p-benzoquinone imine: a cytochrome P-450-mediated oxidation product of acetaminophen. Proc. Natl. Acad. Sci. U.S.A. 81, 13271331. (17) Chen, W., Shockor, J. P., Tonge, R., Hunter, A., Gartner, C., and Nelson, S. D. (1999) Protein and nonprotein cysteinyl thiol modification by N-acetyl-p-benzoquinone imine via a novel ipso adduct. Biochemistry 38, 8159-8166. (18) Baillie, T. A., and Slatter, J. G. (1991) Glutathione: a vehicle for the transport of chemically reactive metabolites in vivo. Acc. Chem. Res. 24, 264-270. (19) Hoffmann, K.-J., Streeter, A. J., Axworthy, D. B., and Baillie, T. A. (1985) Identification of the major covalent adduct formed in vitro and in vivo between acetaminophen and mouse liver proteins. Mol. Pharmacol. 27, 566-573. (20) Streeter, A. J., Bjorge, S. M., Axworthy, D. B., Nelson, S. D., and Baillie, T. A. (1984) The microsomal metabolism and site of covalent binding to protein of 3′-hydroxyacetanilide, a nonhepatotoxic positional isomer of acetaminophen. Drug Metab. Dispos. 12, 565-576. (21) Dennehy, M. K., Richards, K. A. M., Wernke, G. R., Shyr, G. R., and Liebler, D. C. (2006) Cytosolic and nuclear protein targets of thiol-reactive electrophiles. Chem. Res. Toxicol. 19, 20-29. (22) Liebler, D. C., and Guengerich, F. P. (2005) Elucidating mechanisms of drug-induced toxicity. Nat. ReV. Drug DiscoVery 4, 410-420. (23) Uetrecht, J. P. (1999) New concepts in immunology relevant to idiosyncratic drug reactions: The “danger hypothesis” and innate
Chem. Res. Toxicol., Vol. 19, No. 7, 2006 893 immune system. Chem. Res. Toxicol. 12, 387-395. (24) Shenton, J. M., Chen, J., and Uetrecht, J. P. (2004) Animal models of idiosyncratic drug reactions. Chem.-Biol. Interact. 150, 53-70. (25) Park, B. K., Kitteringham, N. R., Maggs, J. L., Pirmohamed, M., and Williams, D. P. (2005) The role of metabolic activation in drug-induced hepatotoxicity. Annu. ReV. Pharmacol. Toxicol. 45, 177-202. (26) Walgren, J. L., Mitchell, M. D., and Thompson, D. C. (2005) Role of metabolism in drug-induced idiosyncratic hepatotoxicity. Crit. ReV. Toxicol. 35, 325-361. (27) Zhou, S., Chan, E., Duan, W., Huang, M., and Chen, Y.-Z. (2005) Drug bioactivation, covalent binding to target proteins and toxicity relevance. Drug Metab. ReV. 1, 41-213. (28) Kaplowitz, N. (2005) Idiosyncratic drug hepatotoxicity. Nat. ReV. Drug DiscoVery 4, 489-499. (29) Mitchell, J. R., Jollow, D. J., Potter, W. Z., Davis, D. C., Gillette, J. R., and Brodie, B. B. (1973) Acetaminophen-induced hepatic necrosis. I. The role of drug metabolism. J. Pharmacol. Exp. Ther. 187, 185194. (30) Liebler, D. C. (2006) The poisons within. Application of toxicology mechanisms to fundamental disease processes. Chem. Res. Toxicol. 19, 610-613.
TX060062O
